Skip to main content
Glama
GayathriNarayanan1610

MCP Tool Connector

MCP Tool Connector

An MCP server that gives an LLM host (e.g. Claude) two tools: it can search the firm's documents and look up a company via an authenticated API the "connect frontier models to our files and data, with auth" use case.

Demo

Register the server with Claude Desktop, then ask Claude something that triggers search_documents, and something that triggers lookup_company.

Related MCP server: Company MCP Server

What it does

Exposes two tools over the Model Context Protocol:

  • search_documents(query, k) — makes a document folder query-able (no auth).

  • lookup_company(name) — calls an authenticated external API (bearer token), making the API + authentication path explicit.

Authentication (done on purpose)

lookup_company reads DATA_API_TOKEN from the environment (never hardcoded) and returns a clear "authentication required" message when it is absent rather than failing silently. Use a scoped, read-only token; full OAuth is the production step.

Run it

pip install -r requirements.txt
pytest -q                      # tools work + both register with the MCP server
python server.py               # run as an MCP (stdio) server

Then add it to Claude Desktop using claude_desktop_config.example.json (use an absolute path to server.py). Confirm the exact MCP SDK usage against the current docs at modelcontextprotocol.io.

Built with

Python · MCP Python SDK (FastMCP) · httpx

Related MCP Connectors

Related MCP Servers

  • -
    license
    Not graded
    quality
    Not graded
    maintenance
    Enables deep analysis of SEC EDGAR filings through universal company search, document content extraction, and advanced filing search capabilities. Provides AI-ready access to business descriptions, risk factors, financial statements, and full-text search across any public company's SEC documents.
    -
  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI assistants and developer tools to securely access and interact with an organization's enterprise knowledge, documents, and people through natural language while respecting existing access permissions.
    165
    MIT